The Roles of IgG, IgM Rheumatoid Factor, and
their Complexes in the Induction of
Polymorphonuclear Leukocyte Chemotactic
Factor from Complement
Teresa Wagner, … , George Abraham, John Baum
J Clin Invest.
1974;53(6):1503-1511. https://doi.org/10.1172/JCI107700.
The induction of chemotactic factor by rheumatoid factor (RF) and by rheumatoid complexes
and their constituents was investigated. The presence of chemotactic factor was measured
by the number of polymorphonuclear leukocytes (obtained from normal individuals)
attracted through a 3 µm Millipore filter and is expressed as a “chemotactic index.” The
chemotactic index was used to measure the effect of immunoglobulins and their complexes
in stimulating production of complement-derived chemotactic factors from the sera of normal
individuals.
Two sources of IgG (F-II and DEAE-purified IgG) were used. These were prepared for use in
the native and heat-aggregated states. The same chemotactic index was obtained with both
these preparations of IgG. The chemotactic index increases when (
a
) increasing
concentrations of gamma globulin were added to a constant concentration of complement
and (
b
) when the amount of complement alone was increased.
The addition of a purified IgM RF to the mixture of IgG and complement caused a decrease
in the chemotactic index. Incubation of IgG and complement before addition of IgM RF
produced no change in the chemotactic index. The addition of a nonrheumatoid factor IgM
to IgM and complement had no effect on the chemotactic index. IgM RF and IgG alone were
not chemotactic.
These studies confirm the concept of “complement deviation” by IgM RF which occurs with,
and as a result […]
Research Article
Find the latest version:
The Roles of
IgG,
IgM
Rheumatoid
Factor,
and
their
Complexes
in the Induction of
Polymorphonuclear
Leukocyte Chemotactic
Factor from
Complement
TERESA
WAGNER, GEORGE
ABRAHAM,
and
JoHN
BAUM
Fromthe Arthritis andClinicalImmunology Unit,MonroeCommunity Hospital, and the Department of Medicine, University ofRochesterSchoolofMedicine
andDentistry, Rochester,New York14642
ABST R AC T The induction of chemotactic factor
by
rheumatoid factor (RF) and by rheumatoid complexesand theirconstituents was
investigated.
The presence ofchemotactic factor was measuredby the number of poly-morphonuclear leukocytes (obtained from normal
indi-viduals) attracted
through
a3AmMillipore
filter and is expressed as a "chemotactic index." The chemotactic in-dex was used to measure the effect of immunoglobulinsand their complexes in stimulating
production
ofcom-plement-derived chemotactic factors from the sera of normalindividuals.
Two sources of IgG (F-II and
DEAE-purified IgG)
were used. These were prepared for use in the nativeand heat-aggregated states. Thesame chemotactic index
was obtained with both these preparations of
IgG.
The chemotactic index increases when(a) increasing
concen-trations of gamma globulin were added to a constant
concentration of complement and (b) when the amount
of complementalonewasincreased.
The addition ofa
purified
IgM
RF to the mixture ofIgG and complement caused a decrease in the chemotac-tic index.
Incubation
ofIgG
andcomplement
before ad-dition ofIgM
RF produced no change in the chemo-tactic index. The addition of a nonrheumatoid factorIgM to IgM and complement had no effect on the chemotactic index.
IgM
RF andIgG
alone were notchemotactic.
These studies confirm the concept of
"complement
de-viation"
byIgM
RF whichoccurs with, andas a resultThis work was presented in part at the 36th Annual
Meeting of the American Rheumatism Association, Dallas,
Tex., 9June 1972.
Dr. Wagner's present address is: Instytut Rheumatolo-giczny, Warszawa ul, Spartanska 1, Poland.
Received for publication 29 March 1973 and in revised
form 21September1973.
of, the immune complex formation between IgG and IgM RF. The successful activation of complement chem-otactic factors by IgG may explain the synovitis with high polymorphonuclear leukocyte counts found in RF-negative arthritis.
INTRODUCTION
Previous studies have demonstrated that polymorphonu-clear leukocytes (PMN)1 obtained from patients with rheumatoid arthritis exhibited decreased chemotaxis (1). Diminished chemotaxis has also been produced by
incubation of PMN obtained from normal individuals with complexes prepared from purified rheumatoid fac-tor and gamma-G globulin (1). However, even' in those
patients with decreased chemotaxis of their peripheral
PMN, an acutely inflamed rheumatoid arthritic joint contains large numbers of PMN that have actively mi-grated into the joint. In this investigation, an attempt has been made to elucidate some factors which may be
responsible for the attraction of PMN into the synovial
fluid of the patient with rheumatoid arthritis. Since about20% of patients with clinical rheumatoid arthritis
are rheumatoid factor negative, other mechanisms (aside from complement activation by immune complex
forma-tion)
can be present which may induce the chemotaxis of PMN into the joint. In these studies, results are ob-tained with aninvitro system which shows that even in the absence of rheumatoid factor and rheumatoid factorcomplexes, complement may be activated by gamma-G
globulin so that chemotactic factors are produced. Fur-ther, the inhibitory effects of rheumatoid factor on the activation of chemotactic factors will be shown.
'Abbreziations used in this paper: F-IT, fraction II;
MATERIALS
PMN were obtained primarily from a normal population of four or five laboratory personnel in whom a standard chemotactic index has been reproduced on numerous occa-sions. The sera used as a complement source for these studies were principally from two individuals. Serum from a third individual was occasionally used. All were sero-type AB and Rh negative. Blood was obtained in 50-ml
bleedings; serum was immediately removed, stored in 2-ml
portions at -70'C, and utilized within 2 mo.
Gamma
globulin
preparations
FRACTION II
Fraction II (F-II) (Difco Laboratories, Detroit, Mich.)
was used in aggregated and unaggregated forms. Before use, the F-II was dissolved in Hanks' solution (Difco
Lab-oratories) and freed of aggregates by centrifugation at
35,000 rpm for 90 min in a Beckman Model L preparative
ultracentrifuge (Beckman Instruments, Inc., Spinco Div.,
Palo Alto, Calif.). The upper two-thirds of this solution (termed "supernate") was utilized. This material was
initially checked for the presence of aggregates by the
centrifugation procedure with 'SI-labeled F-II, and then
by placing an aliquot of the supernate on a 10-40%o sucrose gradient. The gradient was centrifuged at 35,000 rpm for
15 h and separated into 30 fractions. Radioactivity in the bottom two fractions was considered to consist of aggre-gates. This comprised less than 10%o of the total
radio-activity in the gradient.
Aggregation procedure. A solution of F-II in Hanks'
medium was heated to 630C for 10 min. The resulting floc-culent was precipitated twice with 0.72 M sodium sulfate
(2). The precipitate was resuspended in Hanks' solution to the required concentration. Most aggregation was per-formed in a pH range of 6.8-7.2. In a few experiments aggregationwasperformed atpH 8.0.
Immunoglobulin G. IgG used was DEAE purified (Miles Laboratory, Inc., Miles Research Div., Kankakee, Ill.). Thiswastreatedsimilarlytothe F-II.
RHEUMATOIDFACTOR
This was prepared as previously described (3) by
im-munoabsorption of serum from a patient with rheumatoid arthritis by utilizing bromoacetyl cellulose-human IgG. The
strongly latex-positive material obtained was subjected to
Sephadex G-200 chromatography and shown to consist of
only IgM by immunodiffusion, and of a single component
by isoelectric focusing and analytical ultracentrifugation. NORMALIGM
Normal IgM was prepared by chromatography of normal human serum on Sephadex G-200 (Pharmacia Fine Chemi-cals, Inc., Piscataway, N. J.). The first peak eluted was
then further purified by DEAE column chromatography by utilizing a linear 0.015-0.15 M, pH 7.35, phosphate buffer
gradient. A latex agglutination test for anti-IgG activity
of the IgM obtained was negative. The material utilized contained no IgG or IgA by immunodiffusion.
Complement
sources
The AB Rh-negative sera used as the source of
com-plement were utilized at dilutions of 1, 10, 20, and 33A% in Hanks' solution. In these sera, IgG was present at
con-centrations of 940 and 1240 mg/100 ml of serum,
respec-tively. The "concentration of complement" in this study
referstotheserumdilution.
METHODS
In each set of experiments, two to four chambers were used for each test situation and each control. Controls
con-sisted of chambers with either sera (without added immuno-alone.
globulin or complexes) or immunoglobulin preparations The method for inducing chemotaxis of PMN, previ-ously described in detail (1, 4), is based on the method of Boyden (5) and is similar to the procedure described by Ward, Cochrane, and Muller-Eberhard (6).
Blood is obtained from normal individuals in 10-ml Vacu-tainer tubes (Becton-Dickinson and Co., Rutherford, N.J.) containing heparin. To this is added 0.75 ml of a 2%o solution of methyl cellulose in normal saline (7). After standing at room temperature for 30-45 min, the plasma layer is removed. The now enriched concentration of PMN is spun down in a Shandon Cytocentrifuge (Shandon
Southern Instruments Inc., Sewickley, Pa.) and directly
deposited on a 3-,um Millipore filter (Millipore Corp.,
Bed-ford, Mass.). The filter is immediately removed and placed
into a modified Sykes-Moore tissue culture chamber (Bellco Glass, Inc., Vineland, N. J.). Through ports in the chamber,
solutions may be injected into the upper (starting side for cell migration) or lower compartments, which are sepa-rated by the filter. Hanks' solution is usually placed in the upper compartment with the PMN. In one or two experi-ments which will be noted, serum (as a source of com-plement) was placed in the upper compartment as a con-trol. The bottom compartment was filled with the test solutions which will be described. After filling, the cham-ber was incubated for 3 h at 37'C. After incubation, the filters were removed and stained by Boyden's method (5), trimmed, and mounted on microscope slides. The cells on both sides of the Millipore filter were counted by the use of a 1-cm square ocular grid divided into 2X2-mm boxes
(8).
The chemotactic index is the number of cells counted in five of the 2-mm squares on the attractant side of the filter, divided by the number of cells on the starting side in a single 2-mm square. The dividend obtained is multi-plied by 500 to produce the chemotactic index. The fields counted were selected to avoid any areas which showed clumping of cells. 10 fields are counted on each filter.
F-II was initially used at concentrations ranging from 0.01 ,ug/ml to 10 mg/ml. Subsequent experiments were performed at concentrations of F-II and IgG from 0.001 to0.1 mg/ml.
Control experiments were performed with inactivated sera (560C for 30 min) to which the aggregated or un-aggregated F-II had been added.
Experiments were performed which compared the chemo-taxis induced by aggregated and unaggregated F-II to that obtained with DEAE column-purified aggregated or un-aggregated IgG.
Experiments were performed with prior heating of the aggregated F-II to 650C for 20 min. The F-II was sub-sequently precipitated by the sodium sulfate method (2). This was to destroy any possible enzyme activity which
might have been responsible for activation of chemotactic factors from complement. In similar experiments with
unaggregated F-II, the material was heated at 56°C for 30 min. The supernate of the centrifuged solution was used.
To rule out serum factors on the surface of the PMN,
several experiments were performed with the PMN washed
three times in Hanks' solution before centrifugation onto theMillipore filter.
F-II and IgG were combined in some experiments with IgM rheumatoid factor and IgM free of rheumatoid fac-tor activity. In all of these experiments, equal concentra-tions of IgG and IgM were used. The concentration ranges were0.01-0.1 mg/ml.
In order to investigate the deviation or inactivation of chemotactic factor, several experiments were performed with prior incubation of the serum (complement), gamma globulin and IgM rheumatoid factor in the following com-binations: (a) F-II plus complement was incubated at 370C
for 30 min. Then rheumatoid factor was added and the mixture placed in the lower compartment of the migration
chamber for the 3-h incubation period; and (b)
rheuma-toid factor plus complement was incubated at 370C for 30 min. Then F-II was added and the mixture placed in the lower compartment for the 3-h incubation period.
RESULTS
Initially the effects of aggregated and nonaggregated
gamma-G-globulin (F-II) were compared at varying concentrations. The results are seen in Table I. The concentration of the complement was held at
33j%.
Comparisonsof the chemotactic values obtained by using concentrations of F-II from 10 to 0.00001 mg/ml dem-onstrate no significant differences when unaggregated oraggregated materialwas placed in the lower chamber with complement. Studies performed with 33A% com-plement alone yieldeda chemotactic index of
363.1±47.1.
Since, as notedabove, the IgG content of the serum Used as a complement source was approximately 10 mg/ml, a concentration of added F-IIof 1 mg/ml thus onlyraises the total concentration of IgG in the chamber about20%.
The relationship of the chemotactic index to the con-centration of gamma globulin is apparent- when it is seen that chemotactic indices significantly greater than the values for serum alone are not reached until this
TABLE I
Chemotaxis of PMN with Aggregated and Unaggregated F-II
Chemotactic index (meanSEM)
Aggregated Unaggregated
With Without With Without Concentration comple- comple- comple- comple-of F-1I ment* ment ment* ment
mg/mi
10.0 640.74±158.3 - 779.5 4172.7 -1.0 620.9 464.9 - 620.0±464.9 -0.2 336.9±461.2 - 477.0±42.5 -0.1 336.2±49.0 46.8 424.0t36.4 60.0 0.01 419.9±65.3 46.0 389.3436.4 44.4 0.001 338.2-4-59.5 62.8 356.34:49.5 64.9 0.0001 254.44±77.2 38.6 228.6±443.8 62.5 0.00001 55.0±10.8 - 112.5428.6
-*Complement-33j%.
TABLE I I
Chemotaxis of PMN's with Different Concentrations of
ComplementandAggregatedF-Il
Complement Aggregated Chemotactic index concentration F-I1 (mean±SEM)
% mg/mi
33w 0 363.1±-97.1
20 0 262.5-±24.3
10 0 233.5±1=51.5 1 0 48.34-11.7 331 0.01 419.9±-65.3
20 0.01 394.04-127.6
10 0.01 223.8--56.5 1 0.01 132.4±4;17.5
concentration of added F-II (aggregated or
unaggre-gated) was reached.
Further evidence that thechemotactic factors are
pro-duced by activation of complement by the added F-II was provided. In 15 control experiments, F-II at a
con-centration of 0.01 mg/ml was added to inactivated sera (560C for 30 min). The mean chemotaxis was reduced to 86.6±+13.4. Several experiments were performed to rule out the possibility that activity of the unaggregated
gammaglobulin was due to the presence of small amounts ofaggregated material present in the preparations. The
F-II utilized was thus labeled with 'I. After labeling, it was centrifuged at 35,000 rpm and the top two-thirds of the centrifuged preparation removed. A portion was then placed on a sucrose gradient and centrifuged at
35,000 rpm at 4VC for 15 h. The gradient was collected in30 fractionsand each was counted in a gamma counter.
The percentage of radioactivity found in the bottom of the gradient averaged 9.8% of the total amount of
radioactivity. Thus, 90% of the gamma G globulin in
theseexperiments was present as unaggregated material. Since numerous concentrations of F-II were utilized,
this amount of aggregation does not appear to influence
the data, since no significant difference in activity is noted between the two forms of the gamma G globulin
(Table I).
In order to measure the effect of the serum used as
the source of complement, experiments were performed using 33A, 20, 10, and 1% concentrations of serum di-luted in
Hanks'
solution. The results are noted in Table II. Although an overall decrease is noted as the serum concentration is lowered, the most marked drop is seenbetween 33A and 20% and from 10 to 1%. Most of the
studies reported in this paper were performed at serum
concentrations of 33& and 20% and are noted for the
appropriateexperiments.
re-TABLE I II
Chemotaxis of PMNatDifferent Complement Concentrations with DifferentPreparationsof Aggregated and
Unaggregated F-II and IgG
Chemotactic index (mean4SEM)
Comple-ment Aggregated
concen- IgG
tration concentration IgG* F-II1 % mg/mi
33k 0.1 289.0i60.5 336.0±42.0 0.01 295.9±44.8 419.9±65.3 0.001 462.2±4146.0
338.2±59.9
20 0.1 368.0±52.0 305.4±40.4
0.01 270.5±-88.0 -384.0±4127.6
0.001 253.3 47.0 497.2 A244.3 Unaggregated
IgG F-II
33W 0.1 434.6±69.9 424.0442.5 0.01 459.6±95.3 389.3±36.4 0.001
444.4±59.5
356.3±49.5
20 0.1 658.0±194.4 271.0±36.7§
0.01 462.5±99.4 268.0±53.6
0.001 525.8±119.9 301.1±61.5
*
IgG-Miles
Laboratories.F-II-Difco Laboratories.
§P =0.02-all other Pnotsignificant.
lated to the chemotactic results obtained,
comparisons
were made between unaggregated and
aggregated
DEAE-purified IgG and the F-II (Difco Laboratories).
Although the F-IH utilized here is almost entirely
IgG,
wecouldnotruleoutthe effect of contaminationbyother
constituents or of preparative denaturation of the
pro-tein. The results ofthese experiments are seen in Table III.
In these experiments, IgG concentrations from 0.1 to
0.001
mg/ml
in33A
and 20% serum were used.Essen-tially no difference between the two preparations was
noted when the data were analyzed by statistical means.
One group of studies
(20%
complement, 0.1 mg/ml ofunaggregated
F-II and IgG) were significantly differentatthe 0.02 level. This difference between the two
prepa-rations is noted in Table III. It stands as an isolated finding, and we do not feel that this represents a true
difference between the two preparations. Control
ex-periments were
performed
with the IgG alone to ruleout the
possibility
of chemotactic activity from contami-nation by kallekrein in this preparation. At concentra-tions of 0.1-0.001mg/ml
ofIgG,
the chemotacticac-'J. D. Smiley. Personal communication.
tivity was
18±3.7
with aggregatedIgG
and24±4.3
with unaggregatedIgG.
In several experiments, the possibility was tested that the similar effects on the induction of chemotactic fac-tors by both aggregated andunaggregated F-II was due to the method of aggregation. A comparison was thus made of the chemotactic index induced by the addition
of aggregated F-II prepared in three different ways from a single batch of F-II. To destroy enzymes which
might activate the production of chemotactic factors, the aggregated material was first prepared by heating the F-II in Hanks' solution to
650C for
20 min. Thechemotactic index was
595±80.
To check pH effects,some of this aggregated F-II was precipitated twice with 0.72 M Na2SO4 at pH 8.0, while a third group of stud-ies was performed with the sodium sulfate precipitation
at pH of 6.8. In the former, the chemotactic index was
431±89
and in the latterexperiment,
467+98. This was done to test the effects of a mildly alkaline pH on the cells or on the aggregatedF-IH.
The results demonstrate that there was no significant effect with any of the three methods of preparation ofF-IL.
The status of the PMN used in these experiments was next investigated. With the technique utilized here, the cells, after concentration with methylcellulose, are placed directly into the cell centrifuge and spun onto the Milli-pore filter. Excess plasma is removed by a filter paper pad. In this group of experiments, the cells were washed three times with Hanks' solution before placing them in the centrifuge. No differences were noted between un-washed and un-washed cells. This mitigates against the ef-fects of any cell-bound globulin or complement compo-nents from the cell donor affecting the results.
A series of experiments was then performed to look at the activation of chemotactic factors by using the
complex formed from rheumatoid factor and IgG.
Vari-ous experiments were performed which involved prior
FIGURE 1 Interrelationships of IgG, IgM rheumatoid
fac-tor (RF), and normal IgM in the induction of chemo-tactic factorsfromcomplement (C).
incubation of one or both immunoglobulin components with complement and each other. A number of experi-ments were also performed tocompare a nonrheumatoid factor IgM with the IgM rheumatoid factor. In Fig. 1
the data are displayed so that the relationship of the results to the different test conditions can be readily
observed. In the left-hand corner is noted the expected result of the reaction of complement with a concentration of F-II at0.01 mg/ml. When rheumatoid factor is added to the other components, there is a significant drop in the chemotactic index to 162.7. This value (shown in the center) demonstrates the inhibitory effect of rheu-matoid factor on the activation of chemotactic factor. If, however, the F-II and complement are preincubated and the rheumatoid factor then added (center left), a
chemotactic index of 243.7 is obtained. This value is
significantly higher than that reached when the rheuma-toid factor was initially present with the F-II and com-plement. Since activation of complement takes place before addition of the rheumatoid factor, the value
obtained is in the same range as that for the F-II and complement alone.
The deviation of complement (reduction of comple-mentactivation) by the rheumatoid factor is also seen in two further experiments. At the top center of the figure,
results are shown in which rheumatoid factor and com-plement are preincubated and then F-II is added. A chemotactic index of 110.4isobtained. Thisdemonstrates that the rheumatoid factor inhibited the production of
chemotactic factor with deviation of complement
ac-tivation, since addition of theF-IIdidnot induce chemo-taxis. This idea is further substantiated by comparison of the index in this experiment with that shown in the lower right where a similar chemotactic index was ob-tained with rheumatoid factor and complement alone,
i.e., without addition of F-II. This same degree of in-hibition of complement activation is also shown in the experiment in the upper right. Here, the
ability
of the F-II to induce complement activation was blocked byits prior incubation with rheumatoid factor.
The chemotactic indices shown in parentheses in the top and right boxes are those values produced when non-rheumatoid factor IgM was used instead of the rheuma-toid factor. While only small amounts of material were available for a few experiments in each case, it is nevertheless apparent that deviation of complement did
not occur with this material. It further shows that the
effect of the rheumatoid factor appears to be specific
in each case where it reacted with either complement or
F-IH.
In several experiments, the purified rheumatoid factor
was added to the IgG prepared by column
chromatog-raphy. This was done to demonstrate that a
preparation
of
IgG
of greaterpurity
would give results similar tothose seen when utilizing F-IH. The same degree of complement deviation was seen with both preparations.
The chemotactic index of rheumatoid factor plus F-IH
andcomplement was 162.7, while with the purified IgG
with rheumatoid factor and complement, the chemotac-tic index was 155.9. This demonstrates the same degree
ofreactivity of the rheumatoid factor with both prepara-tionsof IgG.
DISCUSSION
The preeminent hypothesis for the cause of the synovitis -of rheumatoid arthritis considers the inflammatory stim-ulus to be the complex formed by rheumatoid factor
(IgG and/or IgM) and gamma globulin (IgG) (9).
The formation of this complex presumably triggers com-plement activation, which results in the production of chemotactic factors (C3a, C5a, C567). These attract PMN into the joint, where they phagocytose the com-plexes (10). The lysosomal enzymes released (11) eventually lead to destruction of the joint. However, this theory does notaccount for those patients in whom rheumatoid factor is not present and in whom synovitis andinflammation can be as severe as that found in rheu-matoid factor-positive patients (12).
Hedberg, in a comparison of synovial fluids from fac-tor-negative and factor-positive patients, showed no
dif-ference in the total white blood cell count of the synovial fluid (13). This was also noted previously by Hollings-worth, Siegel, and Cressey (14), who found the per-centage of PMN to be as high in the synovial fluid of
factor-negative patients as in those patients in whom rheumatoid factor is present. In addition, it has been shown that theIgG concentration in synovial fluids from these two groups is similar and indeed higher than in any other form of arthritis (13, 15). Smiley, Sachs, and
Ziff (16) found the major immunoglobulin produced by explants from synovial tissue was IgG. Therefore,
an-other mechanism, possibly related to the presence of
IgG, might account for the synovitis and large numbers of PMN found in the synovial fluid of patients with
factor-negative rheumatoid arthritis.
Furtherconfirmation of the similarities in thesynovial
fluids from factor-positive and factor-negative patients
is found in the studyby Zvaifler (17) of the breakdown product of C3 in these two groups. To account for the presence of this product, the reaction of aggregated
gamma globulin with rheumatoid factor is postulated.
However, in his studies there was an equally significant number of rheumatoid factor-negative patients in whose
synovial fluids this material was found. These data are
consistent with the experimental results found in our
studies.
The importance of
IgG
in rheumatoid arthritis is(18), who looked at fluorescent staining in the skin and
vessels of patients with rheumatoid arthritis, comparing
them to controls. They found a significant difference in the localization of IgG and IgA in the walls of small
cutaneous vessels. These workers point out that earlier studies also showed IgG localized in the small vessel walls in rheumatoid arthritis in synovial tissue and skin. It should be noted that they found no difference in the
deposition of IgM between the rheumatoid arthritis and controlsubjects.
Restifo,Lussier, Rawson, Rockey, and Hollander (19)
have suggested that inflammation may be produced by
the injections of 7S gamma globulin into an inactive
joint of patients with rheumatoid arthritis. Although
complex formation was felt to be responsible for the production of the synovitis, several interesting points
are noted in their study. The highest white blood cell count reported was found in a joint injected with
un-aggregated gamma globulin where the synovial fluid had the lowest latex titer (1: 40) of any in the study. Subsequent
investigations
by Hollander, Fudenberg,Rawson, Abelson, and Torralba (20) confirmed the
marked effect of unaggregated gamma globulin from rheumatoid arthritis patients on the induction of an
in-flammatory reaction in the joints of rheumatoid factor-positive patients. In one experiment in which nonrheu-matoid gamma globulin was used, the effect of the
un-aggregatedgamma
globulin
wasslightly better in induc-ing the cellular response. These in vivo studies support the biological activity of unaggregated gamma globulinwehave found in our invitro experiments. It should be noted thatSliwinski andZvaifler (21) in similar
experi-ments were unable to reproduce joint inflammation with the injection of aggregated gamma globulin, althoughone
patient
responded to the injection of unaggregatedgamma globulin with a marked inflammatory response
(which
wasnotreproducible).Aggregate-free IgG has been shown to fix comple-ment. Runge and Mills (22) showed that aggregate-free gamma
globulin
fixed complement in 5 of 11 patients,and the fixation found in several was up to
30%.
No difference was found between native or aggregatedau-tologous IgGwhen they performed skin tests in rheuma-toid arthritis
patients. Augener, Grey, Cooper,
andMuller-Eberhard atLa
Jolla
have also demonstrated the binding of monomericIgG
with Cl. This binding wasless than that which took place after aggregation of the
same native
subgroup
ofIgG (23).
Normansell (24) found no convincing evidence that
rheumatoid factor reacts more strongly with altered
gamma G globulin than with native gamma G globulin.
It is clear that the results we have obtained, which show
no
difference
betweenheat-aggregated
gamma Gglobu-lin and unaltered gamma G
globulin,
are akin to asub-sequent demonstration by Normansell
(25)
that the binding of rheumatoid factor to heat-aggregated gammaG globulin was not significantly stronger than that to
unaltered gamma G globulin. Schur and Kunkel
(26)
also found that rheumatoid factor would react withun-aggregated gamma globulin. Our results are consistent
with the data in these studies (Fig.
1).
In the work of Hurd, LoSpalluto, and Ziff (27), it was feltby the authors that the inclusions found in
nor-mal PMN after the addition of IgM rheumatoid factor to rheumatoid factor-negative synovial fluids was due
to the reaction of the added rheumatoid factor with al-tered gamma globulin present in the synovial fluid.
However, they could not otherwise demonstrate the presence of aggregated gamma globulin. Since as
indi-cated above, Normansell
(25)
and Schur and Kunkel(26) showed rheumatoid factor could react with unag-gregated gamma globulin, there may not have to be a
requirement for aggregation of the gamma globulin in thesynovial fluid to produce the complexes.
Numerous workers have found this same phenomenon in varying degrees. Allen and Kunkel (28) called atten-tion to an autospecific rheumatoid factor which had strong affinity for native gamma globulin. They also
re-ported this in gamma myeloma proteins. Oreskes and
Shore (29) showed that there was an inhibitory effect
of added native gamma globulin on agglutination titers. Although the effect was not as marked as with the same gamma globulin aggregated by heating, there was a
sig-nificant effect with the native gamma globulin. Han-nestad showed in a number of experiments, that native gamma globulin was firmly bound to rheumatoid factor and demonstrated clearly that this was not aggregated gamma globulin (30). He felt that when this reactive complex was present in sera, it represented the hidden rheumatoid factors described by Allen and Kunkel (28). The complex could be inhibited by low concentrations of native gamma globulin. In these studies he was able
to show that 7S gamma globulin combined firmly with
isolatedgamma M rheumatoid factor. This complex was
accompanied by similar sedimentationproperties and the
same inhibition of rheumatoid factor activity as was noted in whole sera. Labeled 7S gamma globulin was shown in sucrose density studies to be found with the
macroglobulin rheumatoid factor.
The similarity of effects with both the unaggregated and aggregated gamma globulins and their reactions
with the rheumatoid factor used in this experiment has also been demonstrated by Pike and Schultz (31). They studied the inhibitory effect of several types of gamma globulin on the titer or rheumatoid arthritis sera in the sensitized sheep cell test. They found in one rheumatoid arthritis sera the same 16-fold decrease in titer with un-heated human gamma globulin as with human gamma
globulin that had been heated at 630C for 10 min. The results seen in our experiment are consistent with their
findings and would indicate that the similarity of results with aggregated and unaggregated gamma globulin may
be more likely due to the particular rheumatoid factor used in the studies. This was also shown by Allen and Kunkel (28) and indicates our rheumatoid factor prob-ably represents a factor with a high specificity for na-tive unaltered gamma G globulin.
In our in vitro experiments we have shown evidence that leads to a hypothesis for the role of IgG as the principle factor responsible for the chemotaxis of PMN into the joint of the rheumatoid factor-negative, as well as the rheumatoid factor-positive, patients. Indeed it appears that rheumatoid factor, by preventing the ac-tivation of chemotactic factors, may be playing a pro-tective role. This blocking effect of rheumatoid factor
on complement activation has been pointed out by sev-eral groups (32-34). Schmid, Roitt, and Rocha (32) related this effect to the concentration of IgG in the reaction system. Our control experiments with nonrheu-matoid factor IgM show this inhibition to be a property of the rheumatoid factor activity of the macroglobulin. This "deviation" of complement by rheumatoid factor
wasavoided by prior incubation of IgG with complement
(Fig.
1).
Evidence for the blocking activity of IgMantibody in complement fixationhas been shown for other IgM
anti-bodies beside rheumatoid factor. Scott and Russell (35)
recently demonstrated a similar type of blocking of
complement fixation with antidengue IgM antibody. The blocking effect of rheumatoid factor has also been
found in asystem wherethrombocyte isoantibodies were
blockedby theaddition of rheumatoid factor (36). In early rheumatoid arthritis there is apparently little relationship between the presence of large numbers ofPMN and rheumatoid factor in synovialfluid. In fact, inastudybySchumacher and Kitridou
(37),
the patient with the highest titer of rheumatoid factor in the syno-vial fluid had one of the lower leukocyte counts, with a low percentage of PMN. In the two patients with thehighest leukocytecount
(15,000
and20,000),
onehadnorheumatoid factor and the other had a minimal amount.
Both of these patients also had the highest percentage of PMN. This is of interest in the view of our finding
that in the presence of rheumatoid factor, we found a
decrease in chemotaxis (Fig. 1).
Since we found that adding IgG (increasing the con-centration in the chamberby
20%
or more) producedan increase in chemotaxis, it is apparent that the number of PMN in thejoints
of patients with rheumatoidar-thritis may have some relationship to the high levels of gamma globulin found in the synovial fluid of patients
with this disease, whether rheumatoid factor is present
or absent.
Further support for the importance of gamma
globu-lin in joint inflammation comes from the study of
Cham-berlain, Shapland, and Roitt (38), who showed that autologous heat-aggregated IgG, injected intradermally
in patients with rheumatoid arthritis and in controls,
produces an "Arthus-type" reaction in the patients.
However, their studies differ from our results because the skin reaction they produced with heat-aggregated
material was significantly greater than with unaggre-gated IgG.
The chemotactic activity in serum alone (which serves as the source of complement) is probably due to the presence ofIgG reaction with the complement. Albumin has been 'shown not to be chemotactic (39). Kinsella, Baum, and Ziff (40) found morphological evidence for the association of IgG and complement factors in their appearance in a diffuse cytoplasmic staining pattern in the A and C cells of synovial tissue. The same degree of
fluorescent staining was seen in rheumatoid factor-posi-tive and rheumatoid factor-negafactor-posi-tive patients.Evenin cells from rheumatoid factor-positive synovial tissue, this as-sociation was found without staining for IgG and IgM
rheumatoid factors. This was confirmed by the studies of
Bonomo, Tursi, Trizio, Gillardi, and Dammacco, (41) who found similar staining patterns for IgG and com-plement in synovial cells.
In a recent article on the pathogenesis of joint in-flammation in rheumatoid arthritis, Zvaifler (42) postu-lated the activation of chemotactic factors by immune complexes to account for the "remarkable accumulation of neutrophiles in rheumatoid synovial effusions.. On the basis of our investigation, we believe the induc-tion of chemotactic factors from complement can occur
in the joint in the absence of complex formation. We
feel that the clinical status of the joint inflammation in rheumatoid factor-negative patients with accumulations of PMN like thatseen in factor-positive patients is more related to the similarly elevated levels of IgG found in theirsynovial fluids. However, the presence of complexes and their ingestion bythe PMN and phagocyticsynovial cells causes an outpouring of lysosomal enzymes into the synovial fluid. This, then, can explain the well-known difference in the pathophysiology between these
two groups of patients-the factor-positive
complex-forming group of patients show more erosive disease (43). Even if the IgG is altered or aggregated in the rheumatoid factor-negative patients, it apparently will have little or no effect on the release of lysosomal
en-zymes from PMN.
From these data, it seems probable that gamma
percent of large numbers of leukocytes in the synovial fluid. It has been well demonstrated that factor-negative rheumatoid arthritis does not lead to the degree of joint destruction seen with factor-positive arthritis. It has also been shown that lysosomal enzyme titers are lower in the synovial fluid in factor-negative arthritis (44). This would indicate that joint damage is more likely to appear when complexes can be ingested and cause leak-age of lysosomal enzymes into the synovial fluid. The mere presence of thePMN in high concentrations thus is not sufficient to lead to high lysosomal enzyme levels and joint destruction. On the basis of our studies, it is apparent that rheumatoid factor complex formation is not necessary for theinflux of PMN into the joint, but may theoretically inhibit this process to some extent.
ACKNOWLEDGMENTS
We are indebted to Dr. Peter Ward for his advice during
the course of this study. We wish to thank Mts. Charene
Winney for expert technical assistance and Mrs. Maureen E. Laffin for her skillful preparation of this manuscript.
'This study was supported in part by a grant from the Monroe County Chapter of the Arthritis Foundation.
REFERENCES
1. Mowat, A. G., and J. Baum. 1971. Chemotaxis of poly-morphonuclear leukocytes from patients with rheuma-toid arthritis. J. Clin. Invest. 50: 2541.
2. Christian, C. L. 195& Characterization of the "re-actant" (gamma globulin factor) in the F II precipitin
reaction and the F II tanned sheep cell agglutination test. J.Exp. Med. 108: 139.
3. Abraham, G. N., R. A. Clark, and J. H. Vaughan.
1972. Characterization of an IgA rheumatoid factor: binding properties and reactivity with the subclasses of human -yG globulin. Immunochemistry. 9: 301. 4. Baum, J., A. G. Mowat, and J. A. Kirk. 1971. A
sim-plified method for the measurement of chemotaxis of
polymorphonuclear leukocytes from human blood.- J. Lab. Clin. Med. 77: 501.
5. Boyden, S. 1962. The chemotactic effect of mixtures of antibody and antigen on polymorphonuclear leuko-cytes. J.Exp. Med. 115:453.
6. Ward, P. A., C. G. Cochrane, and H. J. Muller-Eber-hard. 1966.
Further
studies on the chemotactic factor of complement and its formation in vivo. Immunology. 11: 141.7. Boyum, A. 1968. Separation of leukocytes from blood and bone marrow. A two-phase system for removal of red cells with methylcellulose as erythrocyte-aggregat-ing agent. Scand. J. Clin. Lab. Invest. Suppl. 21: 97. 8. Zivkovic, M., and J. Baum. 1972. Chemotaxis of
poly-morphonuclear leukocytes from patients with systemic lupus erythematosus and Felty's syndrome. Immunol.
Commun. 1: 39.
9. Zvaifler, N. J. 1970. Further speculation of the
patho-genesis of joint inflammation in rheumatoid arthritis. Arthritis Rheum. 13: 895.
10. Hurd, E. R., T. D. Kinsella, and M. Ziff. 1971.
Im-munohistologic studies of synoviocytes and synovial exudate cells. J. Exp. Med. 134(Pt. 2): 296 s.
11. Hawkins, D., and S. Peeters. 1971. The response of
polymorphonuclear leukocytes to immune complexes in vitro. Lab. Invest. 24: 483.
12. Nesterov, A. J., and A. G. Astapenko. 1967. Aspects clinico-immunologiques de l'arthrite rheumatoide
sero-negative. European Congress for Rheumatology. Lis-bon. October 1967.
13. Hedberg, H. 1967. Studies on synovial fluid in arthritis. Acta. Med.Scand.Suppl. 479: 36.
14. Hollingsworth, J. W., E. R. Siegel, and W. A. Cressey.
1967. Granulocyte survival in synovial exudate of pa-tients with rheumatoid arthritis and other inflammatory
jointdiseases. Yale J. Biol. Med. 39: 289.
15. Hrncirf, Z., M. Tichy, M. Salavec, and J. Vazrina. 1972. Immunoglobulins A, G, and M in synovial fluid in rheu-matoid arthritis. Reactive synovitis of local origin and in post mortem synovial fluid. Ann. Rheum. Dis. 31: 325.
16. Smiley, J. D., C. Sachs, and M. Ziff. 1968. In vitro synthesis of immunoglobulin by rheumatoid synovial membrane. J.Clin. Invest. 47:624.
17. Zvaifler, N. J. 1969. Breakdown products of C'3 in human synovial fluids. J. Clin. Invest. 48: 1532. 18. Larsson, O., and F. Lithner. 1972. Localization of
vari-ous plasma proteins in the skin in rheumatoid arthritis. Acta. Med.Scand. 192: 13.
19. Restifo, R., A. J. Lussier, A. J. Rawson, J. H. Rockey, and J. L. Hollander. 1965. Studies on the pathogenesis of rheumatoid joint inflammation. III. The experimental production of arthritis by the intra-articular injection ofpurified 7S gammaglobulin. Ann. Int. Med. 62: 285. 20. Hollander, J. C., H. H. Fudenberg, A. J. Rawson, N.
M. Abelson, and T. P. Torralba. 1966. Further studies on the pathogenesis of rheumatoid joint inflammation. Arthritis Rheum. 9:675.
21. Sliwinski, A. J., and N. J. Zvaifler. 1969. The removal of aggregated and nonaggregated autologous gamma globulin from rheumatoid joints. Arthritis Rheum. 12: 504.
22. Runge, L. A., and J. A. Mills. 1971. Invivo and invitro
responses to autologous IgG in patients with rheuma-toidarthritis. Arthritis Rheum. 14: 631.
23. Augener, W., H. M. Grey, N. R. Cooper, and H. J.
Muller-Eberhard. 1971. The reaction of monomeric
and aggregated immunoglobulins with C1.
Immuno-chemistry.8: 1011.
24. Normansell, D. E. 1970. Anti-y-globulins in rheumatoid arthritis sera-I. Studies on the 22S complex.
Immuno-chemistry.7:787.
25. Normansell, D. E. 1971. Anti--y-globulin in rheumatoid arthritis sera-II. The reactivity of anti-yglobulin
rheu-matoid factors with altered yG-globulin.
Immunochem-istry. 8: 593.
26. Schur, P. H., and H. G. Kunkel. 1965. The reactivity
of 19S anti-gamma globulins with native 7S gamma globulin. Arthritis Rheum. 8: 468.
27. Hurd, E. R., J. LoSpalluto, and M. Ziff. 1970. Forma-tion of leukocyte inductions in normal
polymorpho-nuclear cells incubated with synovial fluid. Arthritis Rheum. 13:724.
28. Allen, J. C., and H. G. Kunkel. 1964. Rheumatoid fac-tors with hidden autospecificity. Arthritis Rheum. 7:
725.
29. Oreskes, I., and B. Shore. 1966. Dilution dependent autoinhibition of agglutination in rheumatoid arthritis sera and rabbit anti-human yG antisera. J. Immunol. 97: 600.
30. Hannestad, K. 1968. Rheumatoid factors reacting with autologous native 'yG-globulin and joint fluid yG aggre-gates. Clin. Exp. ImmunoL. 3: 671.
31. Pike, R. M., and M. D. Schultz. 1963. Species
speci-ficity of globulins in reactions with rheumatoid ar-thritis serum. J. Immunol. 91: 740.
32. Schmid, F. R., I. M. Roitt, and M. J. Rocha. 1970.
Complement fixation by a two-component antibody sys-tem: immunoglobulin G and immunoglobulin M anti-globulin (rheumatoid factor). Paradoxical effect re-lated to immunoglobulin G concentration. J. Exp. Med. 132:673.
33. Romeyn, J. A., and D. M. Bowman. 1967. Inhibition of complementary lysis by rheumatoid sera. Nature (Lond.). 216: 180.
34. Davis, J. S., IV, and A. J. Bollet. 1964. Protection of a complement-sensitive enzyme system by rheumatoid factor. J. Immunol. 92: 139.
35. Scott, R. M., and P. K. Russell. 1972. Complement fix-ation blocking M antibody. Activity of anti-dengue IgM. J. Immunol. 109: 875.
36. Kissmeyer-Nielsen, F., A. Svejgaard, K. E. Kjerbye, and T. Christensen. 1967. Blocking effect of rheuma-toid factor on complement fixation by thrombocyte
iso-antibodies. Vox Sang. 12: 199.
37. Schumacher, H. R., and R. C. Kitridou. 1972. Synovitis
of recent onset. A clinicopathologic study during the first month of disease. Arthritis Rheum. 15: 465. 38. Chamberlain, M. A., C. G. Shapland, and I. M. Roitt.
1970. Response to autologous immunoglobulin G in patients with rheumatoid arthritis. Am. Rheum. Dis. 29: 173.
39. Wilkinson, P. C., I. C. McKay. 1971. Chemotactic
ac-tivityof native anddenatured serumalbumin. Int. Arch. Allergy Appl. Immunol. 41: 237.
40. Kinsella, T. D., J. Baum, and M. Ziff. 1969. Immuno-fluorescent demonstration of an IgG-Pic complex in
synovial lining cells of rheumatoid synovial membrane.
Clin.Exp.Immunol.4:265.
41. Bonomo, L., A. Tursi, D. Trizio, V. Gillardi, and F. Dammacco. 1970. Immune complexes in rheumatoid
synovitis; a mixed staining immunofluorescence study. Immunology. 18:557.
42. Zvaifler, N. J. 1970. Further speculation of the patho-genesis of joint inflammation in rheumatoid arthritis.
ArthritisRheum. 1.3:895.
43. Cats, A., and H. M. Hazevoet. 1970.. Significance of positive tests for rheumatoid factor in the prognosis of rheumatoid arthritis. A follow-up study. Ann. Rheum. Dis. 29: 254.
44. Jasani, M. K., M. Katori, and G. P. Lewis. 1969. Intra-cellular enzymes and kinin enzymes in synovial fluid in